On the epistemology of coincidental events